Prognostic risk factors for early death in patients with acute myeloid leukemia

Introduction: Early death is a known complication in the treatment of patients suffering from acute myeloid leukemia (AML). It has been defined as death occurring within 28 days of the initiation of induction chemotherapy.

Aim: Determining the clinical characteristics of the patient, the risk factors, the frequency, and the most common causes of early death in AML patients.

Materials and methods: This retrospective study included 248 patients with diagnosed AML. At diagnosis, the following demographic and clinical-laboratory characteristics were recorded: sex, age, general functional status, i.e., performance status according to the ECOG scale, complete blood count, LDH level in the blood, percentage of blasts in peripheral blood and bone marrow, fibrinogen, PT, aPTT, D-dimer, and BMI. Patients were treated with induction and reduction chemotherapy and palliative therapy. Statistical analysis was performed using the data from the discharge summaries taken from the registers of the Clinic for Hematology of the Clinical Center of Serbia.

Results: Early death occurred in 53 (21.4%) patients. The prognostic risk factors for early death were the following: age (p = 0.047), ECOG ≥ 2 (p = 0.001), leukocyte count ≥ 30 x 10⁹ /l (p = 0.022), LDH level ≥ 450 U/l (p = 0.022), the percentage of blasts in peripheral blood (p = 0.005) and the percentage of blasts in bone marrow (p = 0.003), PT (p < 0.001), as well as the ISTH score (p = 0.018). The most common cause of early death in patients aged 40 – 65 years was sepsis, while in patients older than 65 years it was respiratory failure.

Conclusion: The study showed that age, the leukocyte count, the percentage of peripheral blood and bone marrow blasts, the ECOG score, LDH, PT, and the ISTH score were significant prognostic risk factors of early death in patients with AML, and that the most common cause of early death in patients aged 40 – 65 years was sepsis, while in patients older than 65 years it was respiratory failure.

INTRODUCTION

Acute myeloid leukemia (AML) is a heterogenous clonal malignant disease of the hematopoietic tissue wherein immature hematopoietic cells proliferate and accumulate in the bone marrow, peripheral blood, and other tissues [1].

AML is a rare malignant disease of hematopoietic tissue, which mostly develops in patients of older age [2]. It belongs to a group of diseases with clearly defined cytogenetic and immunological characteristics, which have previously been published elsewhere [3].

Within the first month of diagnosis and treatment initiation, AML patients may suffer fatal complications. This period is connected with high mortality, most frequently as the result of one of many possible treatment complications [4]. Early death, death in aplasia, i.e., treatment-related mortality (TRM) is a well-known treatment complication in AML patients. It is defined as death occurring within 28 days of the initiation of induction chemotherapy [4]. Infection and hemorrhage connected to cytopenia have been stated as the most frequent causes, however, in most cases it is difficult to predict and determine the cause, based on clinical indicators.

Numerous studies have attempted to determine the prognostic risk factors which could be linked to the occurrence of early death in AML patients [5]. Although previous studies have shown that early death occurs more frequently in older patients, it has also been demonstrated that the Eastern Cooperative Oncology Group (ECOG) performance status is more closely linked to the occurrence of early death than age. In addition to age and the ECOG score, other factors, such as: leukocyte count, platelet count, fibrinogen value, hemoglobin level, lactate dehydrogenase (LDH) level, as well as the hematopoietic cell transplantation comorbidity index (HCT-CI), may impact the occurrence of early death [6].

The goals of this paper were to determine the clinical characteristics of patients, prognostic risk factors for early death, the frequency of early death, as well as the most frequent causes of early death in AML patients, who were treated at the Clinic for Hematology of the Clinical Center of Serbia, between January 2016 and December 2019.

MATERIALS AND METHODS

The research was conducted in the form of a retrospective study, based on the database of the Clinic for Hematology of the Clinical Center of Serbia. The study included 248 AML patients, in whom the disease was diagnosed and who were treated in the period January 2016 – December 2019. The diagnosis of AML was established on the basis of the following: clinical presentation, results of the analysis of peripheral blood and bone marrow, the French–American–British (FAB) classification criteria, as well as on the basis of the World Health Organization (WHO) recommendations [7],[8]. In addition to patients with de novo AML, our study group also included 16 (7%) patients who had previously had other hematological diseases (myelodysplastic syndrome and myeloproliferative diseases).

At diagnosis, the following demographic, clinical, and laboratory characteristics of the patients were recorded: sex, age, ECOG performance status [9], complete blood count (hemoglobin, white blood cell count, thrombocyte count, WBC differential), level of lactate dehydrogenase in the blood (which is especially important to be measured in hematological malignancies and solid tumors) [10], percentage of blasts in peripheral blood and in bone marrow, fibrinogen, prothrombin time (PT), activated partial thromboplastin time (aPTT), D-dimer, as well as the body mass index (BMI). The assessment of comorbidities was performed on the basis of the hematopoietic cell transplantation comorbidity index (HCT- CI), which is used in hematopoietic stem cell transplantation [11]. The disseminated intravascular coagulation score (DIC score) was determined according to the recommendations of the International Society on Thrombosis and Hemostasis (ISTH) [12]. The cytogenetic risk level was determined in keeping with the recommendations of European Leukemia Net (ELN) [13].

All patients aged ≤ 60 years were treated with the standard ‘3+7’ induction chemotherapy (CT) consisting of the following: daunorubicin, at a dose of 60 mg/m² , on days (D) 1, 2 and 3, combined with cytarabine, at a dose of 200 mg/m² per day, continuously through intravenous infusion, for a period of 7 days. Patients aged > 60 years were treated, depending on the ECOG score and the HCT-CI index, with reduced intensity CT. Patients with an ECOG ≤ 2 and an HCT-CI <3, received CT according to the ‘3+7 light’ regimen, consisting of the following: daunorubicin, at a dose of 45 mg/m² , on days 1, 2 and 3, combined with cytarabine, at a dose of 100 mg/m² per day, continuously, through intravenous (IV) infusion, for a period of 7 days; while patients with an ECOG >2 and an HCT-CI ≥3, were treated according to the ‘2+5’ regimen, consisting of the following: daunorubicin, at a dose of 30 mg/m² , IV, D1, 3, and cytarabine, at a dose of 100 mg/m² , IV, continuously, D 1 – 5. In older patients with an ECOG >2 and an HCT-CI ≥3, who did not have an adverse cytogenetic risk according to the ELN classification, the following CT was administered: low doses of cytarabine (20 mg, SQ, per 12h, D 1 – 10), and monotherapy with vepesid, amp. 100 mg, D 1 – 5. Palliative treatment consisted of the administration of cytoreductive therapy (litalir, purinethol) and supportive treatment, applied in patients who could not tolerate any kind of aggressive antileukemic therapy. Cytoreductive therapy was applied in patients who had a high WBC count, for the purpose of reducing the tumor mass. Early death was defined as death within 28 days of the initiation of induction chemotherapy [4].

Data was collected from the discharge summaries taken from the registers of the Clinic of Hematology of the Clinical Center of Serbia. Data was presented with descriptive measures: the median (minimum – maximum), absolute and relative numbers. The significance of the difference between the study groups was tested with the Mann-Whitney U test and the chi-squared test. The criterion for statistical significance was p < 0.05. The software program SPSS Statistics 22 (SPSS Inc., Chicago, IL, USA) was used for statistical data processing.

RESULTS

 The study included 248 adult patients suffering from acute myeloid leukemia, who were diagnosed with the disease and treated at the Clinic for Hematology of the Clinical Center of Serbia, between 2016 and 2019. The study group was composed of 129 (52%) men and 119 (48%) women. The average age of the patients was 54 years (range: 18 - 81). In the 18 – 39 age group, there were 46 (18.5%) patients; in the 40 - 54 age group, there were 64 (25.8%) patients; in the 55 – 64 age group, there were 76 (30.6%) patients; while there were 62 (25%) patients in the age group 65 years and above.

On diagnosis, 164 (67.8%) patients had a good performance status – ECOG score < 2, while 78 (32.2%) patients had an ECOG score ≥ 2.

A high HCT-CI index (HCT-CI ≥ 3) on diagnosis was found in 51 (21.3%) patients, while 189 (78.8%) patients had an HCT-CI < 3.

 There were 159 (64.9%) patients with a WBC count < 30 x 10⁹ /l, while there were 86 (35.1%) patients with a WBC count ≥ 30 x 10⁹ /l.

According to the LDH level in the blood, patients were classified into the group with LDH levels < 450 U/l, which included 95 (42.2%) patients, and the group with LDH levels ≥ 450 U/l, which included 130 (57.8%) patients.

According to the body mass index (BMI), there were 50 (32.7%) obese patients with a BMI ≥ 30, while there were 103 (67.3%) patients with a BMI < 30.

There were 232 (93.5%) patients with de novo AML, while there were 16 (6.5%) patients with secondary AML. Of the de novo AML cases, according to the WHO classification of AMLs, there were 221 (95%) unclassified AML patients, 5 (2%) AML patients with signs of myelodysplasia, while there were 6 (2.5%) cases of AML developing after previous treatment, i.e., therapy-related AML (t-AML).

According to the FAB classification of AMLs, the greatest number of patients belonged to the M4 subtype – 75 (34%) patients, followed by the M2 subtype – 57 (26%) patients, the M5 subtype – 46 (21.5%) patients, and the M1 subtype – 24 (10%) patients. The smallest number of patients belonged to the M0 subtype – 19 (8.5%) patients.

According to ELN cytogenetic and molecular risk stratification, the greatest number of patients had intermediate risk – 157 (65.1%) patients, while 11 (4.6%) patients had favorable risk, and 73 (30.3%) patients had adverse risk.

Of the 248 patients included in the study, 125 (50.4%) patients were treated with chemotherapy, 95 (38.3%) patients were treated with reduction chemotherapy, while 28 (11.3%) patients were treated with palliative chemotherapy.

Early death occurred in 53 (21.4%) patients within the study group. Among the patients treated with intensive chemotherapy, early death occurred in 23 (18.4%) cases. Early death occurred in 17 (17.9%) patients treated with reduction therapy. In the group of patients treated with palliative therapy, early death occurred in 13 (46.3%) patients.

The greatest number of patients in our study suffered induction death after the administration of palliative therapy (46.3%). Cytoreductive therapy was applied in 6 patients, with the aim of reducing the tumor mass, while in other patients treated with palliative treatment, cytoreductive therapy was administered due to the fact that these patients were not candidates for the application of induction therapy, owing to other clinical characteristics (age, comorbidities, etc.).

In 53 (21.4%) patients who suffered early death, the average age was 56.56 years (range: 18 – 71 years). In the 18 – 39 age group, early death occurred in 4 (7.5%) patients; in the 40 – 54 age group, early death occurred in 14 (26.4%) patients; in the 55 – 64 age group, early death occurred in 20 (37.7%) patients; while in the group of patients aged over 65, early death occurred in 15 (28.3%) patients.

The clinical parameters that were studied as prognostic risk factors for early death are presented in Table1.

Table 1. Prognostic risk factors for early death

ED – early death; n – number; ECOG – Eastern Cooperative Oncology Group; HCT-CI – hematopoietic cell transplantation comorbidity index; Le – leukocytes; Hb – hemoglobin; Plt – platelets; PB – peripheral blood; BM – bone marrow; BMI – body mass index ; PT – prothrombin time; PTT – partial thromboplastin time; ISTH – International Society on Thrombosis and Hemostasis of the European Leukemia Net (ELN)

Univariate analysis showed the following significant prognostic risk factors for early death in AML patients: age, the ECOG score, the WBC count, the LDH level, the percentage of blasts in peripheral blood, the percentage of blasts in bone marrow, the PT value, and the ISTH score. Age proved to be a significant prognostic risk factor for early death, as the average age of the patients in whom early death occurred was 56.56 years (p = 0.047). There was a statistically higher probability of induction death occurring in patients who had an ECOG PS ≥ 2, than in patients with an ECOG PS < 2 (p = 0.001). Also, patients with a WBC count ≥ 30 x 10⁹ /l had a higher probability of suffering induction death as compared to patients with a WBC count < 30 x 10⁹ /l (p = 0.022). The LDH blood level ≥ 450 U/l was also shown to be a prognostic risk factor for early death, as compared to LDH < 450 U/l (p = 0.029). The percentage of blasts in peripheral blood (p = 0.005), as well as the percentage of blasts in bone marrow (p = 0.003), proved to be statistically significant prognostic risk factors for early death.

The PT value also proved to be a prognostic risk factor (p < 0.001). The ISTH score (p = 0.018) was also shown to be a prognostic risk factor for early death (Table 1).

The analysis of the influence of other analyzed parameters, as prognostic risk factors of early death, did not show statistically significant results (Table 1). Of the 53 deceased patients, early death occurred most frequently as the result of the following: sepsis – in 17 (32.1%) patients; respiratory insufficiency – in 13 (24.5%) patients; CNS hemorrhage – in 7 (13.2%) patients; pneumonia – in 5 (9.4%) patients; enterocolitis – in 4 (7.5%) patients; heart failure – in 3 (6.7%) patients; acute renal insufficiency – in 3 (6.7%) patients; while in one (1.9%) patient, early death occurred as the result of GIT hemorrhage (Table 2).

Table 2. The most common causes of early death and their prevalence in age groups

CNS – central nervous system; GIT – gastrointestinal tract; ARF – acute renal failure

The most frequent cause of early death in patients aged 40 – 65 years was sepsis – in 12 (35%) patients, while in patients aged over 65 years, the most frequent cause of early death was respiratory insufficiency – in 6 (40%) patients.

DISCUSSION

Early death is a well-known complication of induction chemotherapy applied in AML patients, which is why numerous studies have been carried out so as to indicate prognostic risk factors that lead to early death [6]. In our study, individual patient characteristics proved to be statistically significant prognostic risk factors for early death.

In our study, early death occurred in 53 (21.4%) patients. Walter et al. state that, in their study, early death occurred in 10.4% of patients [4].

In our study, the most frequent cause of early death, in patients aged 40 – 65 years, was sepsis, while in patients older than 65 years, the most common cause of induction death was respiratory insufficiency. As opposed to our study, in a study by Ho et al., it has been indicated that, in their study group, early death in younger patients was more frequently caused by respiratory insufficiency, as compared to the older population in this study group. Also, in their study group, sepsis was not reported amongst the most frequent causes of death [14].

In our study, the average patient age at the time of induction death was 56 years. Our study showed age to be amongst the significant risk factors for early death (p = 0.047). In their study, Othus et al. also state that the patients belonging to the group where induction death occurred were older, with an average age of 66 years [5]. In their study, Ho et al. also indicate that early death occurred more frequently in older patients, however, they did not specify the mean age [14].

The average ECOG score also proved to be a significant prognostic risk factor. Our study also showed that there was a higher percentage of early death in the group of patients with an ECOG PS ≥ 2, as compared to the patients with an ECOG PS < 2 (p = 0.001). In their study, Malkan et al. indicated that the patients with an ECOG score < 2, had a lesser percentage of early mortality, as compared to those whose score was ECOG PS ≥ 2, and they state it as the most significant independent prognostic risk factor for early death [15].

Elevated LDH levels (LDH ≥ 450 U/l) proved to be a significant prognostic factor of early death in the patients involved in our study (p = 0.029). This was also demonstrated by the study carried out by Djunić et al., wherein univariate analysis showed that the rise in the level of LDH in the blood was a significant prognostic risk factor for early death [16].

The statistically significant impact of white blood cell values ≥ 30 x 10⁹ /l on the occurrence of early death in the patients involved in our study has also been demonstrated (p = 0,029). A study by Malkan et al. also showed the influence of elevated leukocyte values on the occurrence of early death in AML patients [15]. On the other hand, in their study, Walter et al. did not report elevated WBC count as a statistically significant prognostic risk factor for early death [4].

In our study, it was shown that that the percentage of blasts in peripheral blood (p = 0.005) as well as the percentage of blasts in bone marrow (p = 0.003) were statistically significant prognostic risk factors. A study by Othus et al. as well as a study by Liu et al., also identified these two parameters as statistically significant prognostic risk factors for induction death [5],[17].

In our study, the value of prothrombin time also proved to be a significant prognostic risk factor for early death (p < 0.001). In our study, an ISTH score ≥ 4 proved to be a significant prognostic risk factor (p = 0.018). A study by Berger et al. indicated the significance of an ISTH score ≥ 4 as a prognostic factor [18]. A study by Libourel et al. also confirmed the significance of these two factors in the development of thrombotic episodes, and consequent induction death [19].

A study by Djunić et al. showed HCT-CI ≥ 3 to be the most significant prognostic risk factor for early death [16]. Also, a study by Ho et al showed HCT-CI ≥ 3 to be a significant prognostic risk factor for early death [14]. However, in our study, HCT-CI ≥ 3 did not prove to be a statistically significant prognostic risk factor for early death (p = 0.523).

CONCLUSION

Our study has shown age, performance status of the patient expressed as the ECOG score, WBC count, LDH level, percentage of blasts in peripheral blood, percentage of blasts in bone marrow, PT value, as well as the ISTH score, to be significant prognostic risk factors for early death in patients with acute myeloid leukemia.

Also, our study has shown the most frequent cause of early death in patients aged 40 – 65 years to be sepsis, while in patients older than 65 years, this has proven to be respiratory insufficiency. In a great number of the cases, it is difficult to determine the cause of early death, based on clinical parameters, which is why these cannot be taken as definitive indicators. Therefore, further research is necessary as well as the application of new diagnostic methods, which would help in the individual assessment of the patient’s status and in the possible prediction and prevention of early death.